Tracking Based Motion Segmentation under Relaxed Statistical Assumptions
King Yuen Wong and Minas E. Spetsakis
Technical Report CS-2003-09
York University
October 14, 2003
Abstract
Many Computer Vision algorithms employ the sum of pixel-wise squared differences between two patches as a statistical measure of similarity. This silently assumes that the noise in every pixel is independent. We present a method that involves a much more general noise model with relaxed independence assumptions but without significant increase in the computational requirements. We apply this technique to the problem of motion segmentation that uses tracking to estimate the motion of each region and then we employ our statistic to classify every pixel as part of a segment or the background. We tested several versions of the algorithm on a variety of image sequences (indoor and outdoor, real and synthetic, constant and varying lighting, stationary and moving camera, one of them with known ground truth) with very good results.
Download paper in PDF format.
The documents distributed by this server have been provided by the contributing authors as a means to ensure timely dissemination of scholarly and technical work on a noncommercial basis. Copyright and all rights therein are maintained by the authors or by other copyright holders, notwithstanding that they have offered their works here electronically. It is understood that all persons copying this information will adhere to the terms and constraints invoked by each author's copyright. These works may not be reposted without the explicit permission of the copyright holder.